R for Sports Analytics Exercises: 20 Practice Problems
Twenty practice problems for sports analytics in R: rankings, ratings, ELO, win probability, player metrics, season summaries. Hidden solutions.
RRun this once before any exercise
library(dplyr)
library(tibble)
library(tidyr)
library(ggplot2)
Exercise 1: Standings from results
Difficulty: Intermediate.
Show solution
RInteractive R
games <- tibble(home = c("A","B","C","A","B"), away = c("B","C","A","C","A"),
home_score = c(2,1,3,0,2), away_score = c(1,1,2,1,2))
games |>
transmute(team = home, points = ifelse(home_score > away_score, 3,
ifelse(home_score == away_score, 1, 0))) |>
bind_rows(games |> transmute(team = away, points = ifelse(away_score > home_score, 3,
ifelse(away_score == home_score, 1, 0)))) |>
group_by(team) |> summarise(pts = sum(points)) |> arrange(desc(pts))
Exercise 2: Win percentage
Difficulty: Beginner.
Show solution
RInteractive R
wins <- 25; losses <- 15
wins / (wins + losses)
Exercise 3: Pythagorean expectation
Difficulty: Advanced.
Show solution
RInteractive R
runs_for <- 720; runs_against <- 650
runs_for^2 / (runs_for^2 + runs_against^2)
Exercise 4: Top scorers
Difficulty: Beginner.
Show solution
RInteractive R
df <- tibble(player = letters[1:5], goals = c(20, 15, 30, 12, 25))
df |> arrange(desc(goals)) |> head(3)
Exercise 5: Z-score for player stat
Difficulty: Intermediate.
Show solution
RInteractive R
goals <- c(20, 15, 30, 12, 25, 18, 22)
(goals - mean(goals)) / sd(goals)
Exercise 6: Goals per game
Difficulty: Beginner.
Show solution
RInteractive R
df <- tibble(player = c("a","b"), goals = c(20, 25), games = c(30, 28))
df |> mutate(g_per_game = goals / games)
Exercise 7: ELO update
Difficulty: Advanced.
Show solution
RInteractive R
elo_update <- function(rA, rB, sA, K = 32) {
eA <- 1 / (1 + 10^((rB - rA) / 400))
rA + K * (sA - eA)
}
elo_update(1500, 1500, 1)
Exercise 8: Apply ELO across season
Difficulty: Advanced.
Show solution
RInteractive R
games <- tibble(team1 = c("A","B","A"), team2 = c("B","C","C"),
winner = c("A","C","A"))
ratings <- c(A = 1500, B = 1500, C = 1500)
for (i in seq_len(nrow(games))) {
r1 <- ratings[games$team1[i]]; r2 <- ratings[games$team2[i]]
s1 <- as.integer(games$winner[i] == games$team1[i])
e1 <- 1 / (1 + 10^((r2 - r1)/400))
ratings[games$team1[i]] <- r1 + 32*(s1 - e1)
ratings[games$team2[i]] <- r2 + 32*((1-s1) - (1-e1))
}
ratings
Exercise 9: Win probability from ELO
Difficulty: Intermediate.
Show solution
RInteractive R
rA <- 1600; rB <- 1500
1 / (1 + 10^((rB - rA) / 400))
Exercise 10: Head-to-head record
Difficulty: Intermediate.
Show solution
RInteractive R
games <- tibble(t1 = c("A","B","A","A"), t2 = c("B","A","B","B"),
winner = c("A","A","B","A"))
games |> filter((t1 == "A" & t2 == "B") | (t1 == "B" & t2 == "A")) |>
count(winner)
Exercise 11: Streak detection
Difficulty: Advanced.
Show solution
RInteractive R
res <- c("W","W","L","W","W","W","L")
rle(res)
Exercise 12: Plus-minus per player
Difficulty: Advanced.
Show solution
RInteractive R
events <- tibble(player = c("a","b","c","a"),
on_court = c(TRUE, TRUE, FALSE, TRUE),
score_change = c(2, 2, 0, -1))
events |> filter(on_court) |> group_by(player) |> summarise(plus_minus = sum(score_change))
Exercise 13: Home-field advantage estimate
Difficulty: Advanced.
Show solution
RInteractive R
games <- tibble(home_pts = c(100, 95, 110, 88, 105),
away_pts = c(95, 92, 100, 90, 102))
mean(games$home_pts - games$away_pts)
Exercise 14: Player percentile
Difficulty: Intermediate.
Show solution
RInteractive R
goals <- c(20, 15, 30, 12, 25, 18, 22, 28)
ecdf(goals)(25)
Exercise 15: Team form (last 5 games)
Difficulty: Intermediate.
Show solution
RInteractive R
games <- tibble(team = rep("A", 10),
result = sample(c("W","L","D"), 10, replace = TRUE),
date = Sys.Date() - 9:0)
games |> arrange(desc(date)) |> head(5) |> count(result)
Exercise 16: Expected goals (xG) aggregate
Difficulty: Advanced.
Show solution
RInteractive R
shots <- tibble(player = c("a","a","b","b","b"),
xg = c(0.1, 0.3, 0.05, 0.4, 0.2))
shots |> group_by(player) |> summarise(total_xg = sum(xg))
Exercise 17: Goals vs xG (efficiency)
Difficulty: Advanced.
Show solution
RInteractive R
players <- tibble(player = c("a","b"), goals = c(5, 3), xg = c(4.2, 5.1))
players |> mutate(over_perform = goals - xg)
Exercise 18: Visualize standings
Difficulty: Intermediate.
Show solution
RInteractive R
standings <- tibble(team = c("A","B","C","D"), pts = c(30, 22, 18, 12))
ggplot2::ggplot(standings, ggplot2::aes(reorder(team, pts), pts)) +
ggplot2::geom_col() + ggplot2::coord_flip()
Exercise 19: Compare positions
Difficulty: Intermediate.
Show solution
RInteractive R
players <- tibble(pos = c("F","M","D","F","M","D"),
goals = c(20, 5, 1, 18, 8, 2))
players |> group_by(pos) |> summarise(mean_goals = mean(goals))
Exercise 20: Days rest between games
Difficulty: Intermediate.
Show solution
RInteractive R
games <- tibble(team = "A", date = as.Date(c("2024-09-01","2024-09-05","2024-09-12")))
games |> mutate(rest = as.integer(date - lag(date)))
What to do next
- EDA-Exercises (shipped), explore your dataset.
- Linear-Regression-Exercises (shipped), model on player stats.